HOW TO REVIEW A CETOL 6σ ANALYSIS Presented by Dan Lange, Manager - Services

HOW TO REVIEW A CETOL 6σ ANALYSIS Presented by Dan Lange, Manager - Services

Abstract Reviewing a CETOL 6σ analysis requires a comprehensive understanding of the assembly being studied and the capabilities of the software. In this presentation we will review a methodical approach used by Sigmetrix engineers to review analyses submitted by newly trained students during post-training support. The approach covers modeling, validation, and results interpretation and can be applied to your own analyses or analyses created by your peer.

KEY SUBJECT MATTER Advisor Window Part order Joint order, type, location, orientation, considerations CETOL 6 Modeler Measurements Model behavior Results CETOL 6 Analyzer Assembly Modeling Create joints Set joint properties Define Objectives Create measurements Set measure properties Review Results Inspect measurement variation plots Inspect sensitivities and contributions Part Modeling Constrain features (dimensions, GD&T, etc.) Specify tolerances Improve Quality/Reduce Cost Adjust tolerance values Incorporate mfg. data Etc. Validate Model Check Advisor messages Show Part Locations Visualize sensitivities Run Analysis Generate Report

Advisor Window Clear all messages Why? Impact can range from no impact to incorrect answer or no answer at all. Action See Advisor Message Table Note: Selecting message in Advisor window will jumps to the item in the CETOL Tree, if applicable CETOL 6σ Advisor Window Unconstrained Part DOF

Advisor Messages ADVISOR MESSAGE "Feature is unconstrained by ##" "All constrained variables are redundant" "The constraint controls no variables" "Measure appears to be discontinuous" "Endpoints invalid for measure type" "Direction feature invalid for this measure" "One or more derivatives failed" "The 2nd order Taylor Series used for the SWC solution has an internal critical point" "The 2nd order Taylor Series used for the SWC solution has a saddle point" ACTION Add dimensional constraint Determine which constraint to keep; delete other Delete constraint if not needed Define measurement direction Chose valid direction feature or delete Review Analyzer Sensitivity List; KB Article - SSC. The measurement has a minimum or maximum that is not at tolerance limit for one or more of the variables. Use visualize sensitivities for the largest contributors to see if any of those variables have internal critical points. Use a scale factor of 1.0 and look for curves in which the maximum or minimum value is not at one of the endpoints of the curve. KB Article - SSC.

Advisor Messages ADVISOR MESSAGE "Component is under-constrained" "Joint is over-constraining" "Object is Invalid" "The GDT callout type is not supported for this feature type "The DRF is not valid for this GDT callout" "The orientation tolerance is larger than allowed by the related location tolerance "The feature is not valid as a datum feature" "The feature is not valid as a multiple datum feature" "No orientation feature for joint with tangent bias" "The tolerance limits are far from the design nominal" "The distribution mean lies outside the tolerance limits" "No orientation feature for joint with ambiguous orientation" ACTION Select part, highlight under-constraint. Add joint Review joint order. Remove DOF Tip: rotational) Yellow triangle; reselect a reference. Workaround or redefine dimension scheme Redefine DRF Adjust orientation or location tolerance. Redefine datum feature Redefine datum feature Define orientation. Review DOFs. Tx-Ty may switch Review tolerance settings for constraint Review distribution. Define orientation. Review DOFs. Tx-Ty may switch

Advisor Messages ADVISOR MESSAGED "Datum feature has no size dimension and is not RFS" "The linked CAD GDT has unsupported attributes" "The CETOL DRF does not match DRF of the linked CAD GDT" "The CAD object linked to this CETOL object is missing" "The dimension nominal does not match the linked CAD dimension's nominal" ACTION This indicates that there is an (M) or (L) modifier applied to a datum feature that does not have a size dimension (e.g., the feature references a Creo datum plane or a datum axis). You do one of the following: 1. Edit the feature and add the related geometry. For example, if the feature references a Creo datum axis, add the hole or shaft geometry to the CETOL feature 2. Remove the the material condition modifying symbol (i.e., change it to RFS) There are some attributes [(P) (F) and (T) or <ST>} that are not supported and their affects are ignored. This happens if you edit the DRF in the CAD system after establishing the link to the CETOL object. You should do one of the following: 1. Delete the object 2. Unlink the object This happens if you link a CETOL dimension to a CAD dimension that may not be directly associated with the geometry that the feature references and that dimension is changed in the CAD model.

Advisor Messages ADVISOR MESSAGE "The tolerance type of the linked CAD dimension is unsupported" "The model has unsupported units" ACTION The dimension has an unsupported tolerance type (e.g., MIN or MAX). I think this is only applicable to SolidWorks models. You get this if you try to synch with a CAD model with unsupported units (e.g., angles in deg/min/sec) Supported linear units: micrometers millimeters centimeters meters mils (1/1000th inch) inches feet angstroms nanometers kilometers micro-inches yards miles Supported angular units: radians degrees milli-radians gradians

Part Order Order parts to represent actual assembly sequence as best as possible Why? Reporting of assembly under-constraints Animations are affected Example DOFs: (6) DOV part versus (4) DOV part Affects solving for closure Action Review order using component highlight

Part Order Matters (6) DOV PART Constrains (4) DOF BASE is ground part. (4) DOF applied to (4) DOV BUSHING results in Over: 0 Under 0 (4) DOV PART BUSHING is ground part. (4) DOF applied to (6) DOV BASE results in Over: 0 Under 2

Joints Order, Location, Type For most conditions*, the first joint controls the orientation of the part. Joints 2, 3, control location. Represent actual assembly process order as best as possible. Check location, orientation & type of joint *Not always shaft held at ends or 3-legged stool Why? Reporting of over-constraints is based on order Animations & results are affected Action: RMB in CETOL Tree Tree Level \ Joints & Measures Show Part Loc TIP: Empty config to show DOF progression

Joints - Considerations Watch for and review the following: Joint balls blurry or have distance between? Not coincident investigate Reduce symbol size Do you want the joint to close the gap or should it be redefined? Custom joints Why was a custom joint used? Does it simplify the analysis? Will it affect answer? Closure, the ability to put the model in the position defined by the joints. Can result in Solved Nominal = CAD nominal. More later

Measurements Verify measurement accurately defines assembly requirement Why? Could lead user to make unnecessary changes, especially nominal design changes Got to ask the right question to get the answer you need Action: Activate a configuration Set assembly in SHOW PART LOC mode Review direction & attachment locations Review nominal values. Plus or minus? Solved nominal = CAD nominal? Should it? Joints used that close gaps? Review Target Limits Fit analysis (1-sided failure limits) or performance analysis Example shows solved nominal different from CAD nominal. CAD constraints centered teeth. Tangent gear teeth altered position of GEAR2.

Configuration Joints Verify assembly position Why? Need to ensure position matches use-case for analysis Action Use SHOW PART LOC Make sure new position of parts does not create interference ANYWHERE in the model. (ProE: Global I/F or Pairs Clearance) Eg. Does it really touch at location 1 first, or does it touch at location 2 first? Focus on parts that move (Continuation from previous example.) BLURRY SMALL CLOSURE JOINT LOCATION EDGES OF GEARS INTERFERE EDGES MADE TANGENT CLOSED

Configuration Joints, continued Example: Does it really touch at location 1 first, or does it touch at location 2 first? Theory was that the connector head orientation was controlled by the contact channel (first image). contact channel head interference Actual contact locations

Model Validation Verify proper movement of model Why? Animations should look natural. If not, sensitivities and results will be incorrect. Joint definition may be incorrect. Dimension scheme may be incorrect E.g. center of width feature versus side Action Run Worst Case Animation Run Sensitivity Animation for top variables, sensitivities and contributions Visualize response to understand how features respond

Model Validation Review Contribution Data Why? Identify tolerance value errors Identify derived variation values Envelop rule Characteristic length Joint scheme may be too simplistic Rotation arbitrarily constrained, eg. planar versus 3 pts Action Access Plots tab and look for outliers Validate any outlier Answer: Are the highest contributors the variables you expected? TOLERANCE TYPO

Model Validation Review Sensitivity Data Why? Gain confidence that measurement, joint and dimension modeling was done properly Action Compare values in sensitivity plot. Do expected variables have sensitivity? Do variables with sensitivity = 0.0 make sense? Size automatic included. Any variables with value order of magnitude different? Don t forget angular variables SMOOTH. NO OUTLIERS UNDERSTAND ZEROS TRANSLATION DON T FORGET ANGLES

Model Validation Review Sensitivity Data, cont. Action Review Sensitivity Graphs Linear when Scale Factor = 1.0? No? consider 2 nd Order NON-LINEAR LINEAR

Model Validation Tip & Guideline Combined Approach Review sensitivity & contribution at same time in Analyzer Action Add New Layout tab in Analyzer Assign Filter

Model Validation Watch measurements during animation Why? Ensure measurement direction & attachment updates consistent with measurement intent, i.e. should measurement location move or stay locked? Action Run worst-case and sensitivity animations Measurement should not jump around other than arrow change If jump occurs during WC animation, at what value? Compare value to statistical result. 16:1 leverage

Model Validation Worst-Case is Conservative Use to id unusual areas of assembly. May be joint DOF not appropriate May be excessive variation limits from characteristic length. Eg. position specification on shallow hole Excessive angle limits? WC

Model Validation Review Variables Why? Ensure inputs have proper distribution and limits Action RMB in CETOL Tree and select Tree Level>Variables Select first variable in list Select Distribution tab Select each variable and see that distribution is as expected

Thank You!! Check Sigmetrix Support Center for Advisor Message List Model Review Checklist